What Will IT Infrastructure Look Like in 10 Years?

By Raul Martynek, DataBank CEO

If we are to believe the latest headlines, it may seem like we’re just a few years away from autonomous cars, powerful AI tools rendering many jobs (and even entire careers) obsolete, and other examples of futuristic technology transforming our daily lives.

While the jury is still out on these use cases and their long-term implications, it is safe to assume we will continue to see the introduction of new technologies as well as the evolution of existing digital tools. In response to these developments, consumers will demand new – and better — products, services, and experiences. Businesses will also react, both to meet these new expectations but also to meet their own goals of becoming more efficient, more agile, and more competitive.

All of this will change current conventional approaches to developing, implementing, and using IT infrastructure. New developments will also force data center providers like DataBank to adapt new strategies and technologies to provide the power, performance, edge presence, and other criteria tomorrow’s businesses will require.

How will IT infrastructure evolve and what will it look like in the next ten years? To answer these questions, it may help to look at three emerging technological trends – AI, 5G, and edge computing – and see how they’re already influencing future generations of IT infrastructure.

AI Workloads: An Exponential Game Changer Requiring Exponentially More Power and Cooling

It’s hard to believe that it’s only been a few months since ChatGPT first became available for public testing in November 2022. Yet, it’s clear that this technology has the ability to transform, revolutionize, and even upend so many of our daily practices and processes. This includes everything from creating content to automating workflows to streamlining administrative tasks and more.

As the use of artificial intelligence continues to explode, AI-driven workloads will require much more processing power, data storage, and network bandwidth. Many companies will have to upgrade some facets of their architecture with high-performance computing, cloud-based data storage, and even new network hardware.

Data center providers supporting these companies will need to provide much more power to a smaller amount of space to support these AI workloads as well as new use cases we may not even be thinking about yet. Consider the fact that ChatGPT currently has nearly 25 million daily visitors, whose search activity translates to approximately 600,000 kilowatt hours per day of consumption – 2,000 times more than what the average U.S. household consumes each day.

DataBank is already seeing high-performance computing cases where our customers now need 30-100 kilowatts for each cabinet in the data center. This is not an incremental increase. These power density requirements are an exponential leap over previous scenarios where data center customers only needed three to 10 kilowatts.

With more power will come the need for more cooling. This means that future data centers will have to rely on new cooling methods – such as rear chilled doors, water to the chip or immersion technologies – as well as other strategies such as more intelligent data center design to provide the right mix of power and cooling.

The Continued Rollout of 5G and the Push to the Mobile Edge

5G has the potential to increase connectivity between people, digital products, and services while delivering higher data speeds, lower latency, increased availability, and other benefits. For businesses and service providers to take full advantage, they need to move beyond hardware-centric, previously centralized infrastructure and implement a more scalable, cloud-native architecture.

Driven by the promise and potential of 5G networks, more and more companies will push their infrastructure and hosted data centers even further to edge markets. This will be true as more companies continue to adopt 5G (and even look to the release of 6G in 2030). This new approach is critical to reach customers at the edge of their network and deliver the enhanced experiences associated with 5G.

As I covered in the sidebar of this article, four critical infrastructure components are needed to manifest the potential of 5G. You can find this at the very end of this article.

In the past, the entire industry leaned toward a centralized model where servers, fiber optic cables, and data storage were consolidated in a relatively small number of “core” data centers locations. With the emergence of 5G and the new era of mobile edge computing, far more data will be created at the edge — near users of the data — and will need to be processed and stored there to support IoT, robotics, gaming, autonomous machinery, and latency-sensitive applications and workloads.

In these mobile edge environments, network providers will increasingly be drawn to Tier 2 and Tier 3 markets, expanded geography within Tier 1 markets and the multi-tenant data centers located in them. They are also expected to take advantage of modular data center platforms that can be deployed anywhere with closer proximity to fiber and cellular tower infrastructure. They will also be able to perform localized internet exchanges in facilities that are more open and neutral than past options.

The Future is Not So Far Away

These trends are already happening now and affecting the way companies are designing their IT architecture. They also raise new questions when it comes to evaluating future data center providers and what options will be right for their business. It’s clear that their infrastructure will have to change to keep pace with industry and technology developments, so getting a proactive head start on strategic planning now – and identifying the right hosting model – can go a long way to ensuring future success.

To learn more about DataBank, and how our we’re designing our data centers to meet tomorrow’s infrastructure needs, please visit www.databank.com today.

###

The 4 Critical Components of Deploying a True Stand-Alone 5G Network:

• Incorporating the full-frequency, three-spectrum-band “layer cake” all the way from cell towers, small cells, or rooftop RF distributions to devices. This weaves together low, mid, and high band frequencies that will provide the Ultra-Wide Broadband (UWB) and Ultra-Low Latency (ULL) of 5G; the carriers are about 25% of the way to completing this component.
• Re-architecting network cores from 4G to 5G virtualized. This will transform the cores from a centralized design to an architecture where a relatively smaller number of control points are disaggregated into dozens of user-plane locations where applications can access the network at a more local level.
• Virtualizing Radio Access Networks (cRAN or vRAN) so cell towers in a region can pool their resources to handle peak user activity and improve handoff between cell zones. This eliminates the need to size each individual macro site for peak activity, which is the case with 4G.
• Leveraging open networks components where the key network functions are ‘virtualized” as software that runs on commodity hardware and allows applications to plug directly into networks. This approach enables a more proximate location for mobile users to access edge computing at cRAN cell tower locations and performing like the internet in terms of aggregation points.